Problem with threading on Solaris

[Martin von Loewis]

bbrox at bbrox.org (Lionel Ulmer) writes:

> By the way, how should I handle all the 'PY_PTHREAD_D*' mess ? 

I'd say ignore it. In the autoconf test, use pthread API as if you'd
have a fully standards compliant system. If that doesn't compile, it
still might if we follow the pthreads draft in use. However, since the
pthreads drafts have been long superceded, anybody using such a system
and wishing to use set_scope will need to supply a patch.

> In some case, we use 'NULL' as the thread attributes, sometimes
> 'pthread_attr_default'. If I do my patch as told, I will have to
> change all the NULL or pthread_attr_default parameters by my own
> variable that will be initialized by 'pthread_attr_init'.

If you think you can work it out, you may try, but my advise is not to.

> Should I change this in all the D4 / D6 / D7 sections or only in the STD
> section (the one used on Solaris) ?

Just change the standard section. I believe it becomes complicated
enough, if you try to continue passing NULL if no attributes are set
(i.e. neither stacksize, nor scope).

Regards,
Martin